Configuring AWS AI
Overview
This guide explains how to configure AWS AI Transcription in PortSIP PBX.
Once configured, PortSIP PBX integrates with Amazon Web Services (AWS) AI services to provide enterprise-grade Speech-to-Text (STT) transcription and optional sentiment analysis for calls and voicemails. These capabilities enhance call analytics, quality management, compliance recording, and overall customer experience.
This guide follows VoIP, UCaaS, and CCaaS best practices and assumes basic familiarity with PortSIP PBX administration and AWS services.
AWS Services Used
PortSIP PBX leverages the following AWS AI services and IAM permissions to enable transcription, analytics, and optional AI capabilities:
AmazonTranscribeFullAccess Provides real-time or post-call Speech-to-Text (STT) transcription for calls and voicemails.
ComprehendFullAccess Provides sentiment analysis and natural language processing (NLP) for transcribed text.
AmazonBedrockFullAccess Enables access to Amazon Bedrock foundation models for advanced AI processing and future AI-powered features.
AmazonS3FullAccess Allows storage and retrieval of transcription artifacts and related data in Amazon S3, if required by the deployment.
Best Practice For production environments, it is recommended to replace full-access policies with custom IAM policies that grant only the minimum permissions required.
Prerequisites
Before you begin, ensure the following requirements are met:
System Administrator privileges in PortSIP PBX
An active AWS account
An AWS IAM user or role with permissions for:
AmazonTranscribeAmazonComprehend(required only if sentiment analysis is enabled)
AWS Access Key ID and Secret Access Key
A selected AWS region where AI services are available
(Optional) An Amazon S3 bucket for storing transcription output, if required by your deployment model
Configuring the PortSIP PBX AI Engine
Step 1: Log in to PortSIP PBX
Sign in to the PortSIP PBX Web Portal using a System Administrator account.
Step 2: Select AWS as the AI Engine
Navigate to menu: Integrations > AI Engine
From the AI Engine drop-down list, select AWS.
This configures PortSIP PBX to use Amazon AI services as the backend for transcription and language analytics.
Step 3: Configure AWS Credentials and Settings
This section describes how to configure AWS credentials, service parameters, and platform API request limits used by PortSIP PBX for AI-powered speech processing features such as Speech-to-Text (STT) and language analysis.
AWS Service Configuration:
Access Key
The AWS Access Key ID associated with an IAM user or role.
This key identifies the AWS account and is required to authenticate requests sent from PortSIP PBX to AWS services.
Secret Key
The AWS Secret Access Key paired with the Access Key ID.
This key is used to cryptographically sign API requests to AWS services and must be kept strictly confidential.
Security Best Practice Use a dedicated IAM user or role with least-privilege permissions, and never use root account credentials.
Region
The AWS region where your AI services are deployed.
Examples include:
us-east-1, us-west-2, eu-central-1, ap-northeast-1
Recommendation Choose the region closest to your PBX deployment or primary user base to minimize latency and improve transcription performance.
Bucket Name
The name of the Amazon S3 bucket used to store audio files, intermediate processing data, or transcription results, depending on your deployment configuration.
Locale
Specifies the language and regional format used for speech recognition and text processing.
Examples include: en-US, en-GB, ja-JP, zh-CN, vi-VN
Language Code
The AWS-specific language code used by speech and language services such as Amazon Transcribe and Amazon Comprehend.
Examples include: de, en, es, pt.
Note This value must match one of the language codes supported by the corresponding AWS service.
AWS Platform API Request Limits:
These settings control how many AI-related requests can be queued and processed concurrently. Proper configuration helps comply with AWS service quotas and protects overall system stability.
Max Speech-to-Text Queue Size
Defines the maximum number of Speech-to-Text requests that can be queued and waiting for processing when the AI service is busy or rate-limited.
Sentiment Analysis Requests
Specifies the maximum number of Sentiment Analysis API requests that PortSIP PBX is allowed to send to AWS.
This limit applies when analyzing transcription text to determine caller sentiment, call quality metrics, or other AI-derived insights.
Step 4: Enable AI Transcription for a Tenant
Log in to PortSIP PBX web portal as a System Administrator.
Navigate to Tenants, select the target tenant, and click Edit.
Select the Features tab and enable AI Transcription.
Navigate to General and enable Enable AI Transcription.
Configure the Daily File Quota to limit the tenant’s daily AI transcription usage and control AWS costs.
Managing AI Transcription Within a Tenant
Once AI transcription is enabled, Tenant Administrators can manage and use AI features as follows:
Log in to the PortSIP PBX web portal as a tenant administrator.
Navigate to Company > General and configure:
Automatically Transcribe Recorded Calls
Automatically Transcribe Voicemails
Navigate to Data Analytics > Call Recordings to view transcription results.
Calls that have already been transcribed display a sentiment indicator.
For calls that have not yet been transcribed, click the transcription icon to manually trigger transcription.
Managing AI Transcription at the User Level
Once AI transcription is enabled for the tenant, Tenant Administrators can manage and use AI transcription features for individual users as follows:
Log in to the PortSIP PBX Web Portal as a Tenant Administrator.
Navigate to Call Manager > Users, then double-click the target user.
Select the Extension tab and enable or disable the following option:
Automatically Transcribe Recorded Calls
Navigate to Data Analytics > Call Recordings to view transcription results.
Calls that have already been transcribed display a sentiment indicator.
For calls that have not yet been transcribed, click the transcription icon to manually trigger transcription.
Operational Notes and Best Practices
Call recording must be enabled for transcription to function.
Only recorded calls and voicemails are eligible for transcription.
Transcription workloads are processed asynchronously and may be queued during peak usage.
Use Daily File Quotas to manage AWS costs and prevent unexpected usage spikes.
For regulated environments, review AWS data residency and retention policies.
Last updated